This article is mentioned in:

Introduction

Sepsis is one of the common causes of death in intensive care unit (ICU) patients. As inflammation spreads into the bloodstream, the deregulated production of cytokines, such as tumor necrosis factor (TNF)-α and interleukin (IL)-1b, triggers high mobility group box (HMGB)1 expression as a late mediator of inflammation (1). Systemic and persistent inflammation may lead to organ dysfunction and ultimately, death (2,3). Despite evidence suggesting that TNF-α, IL-1b and HMGB1 are associated with endotoxin lethality and septic syndrome, unfortunately, monoclonal antibodies against TNF-α and TNF-binding proteins fail to decrease sepsis (2).

Despite the availability of 170 different biomarkers, the complex clinical syndrome of sepsis and related mortality have increased in incidence (4). Microbial infection underlying sepsis accounts for an almost 30% increase in mortality, despite advances in antimicrobial research (5). The majority of studies on sepsis have focused on biomarkers for early diagnosis, treatment or prognosis. Treatment is very unsatisfactory, as early inflammatory cytokines are restored to their basic levels only after sepsis-related death (9).

A more appropriate cytokine to block sepsis includes the late onset inflammatory mediator, HMGB1, a crucial late-phase cytokine in the course of sepsis (6). A high level of HMGB1 is associated with a poor outcome in sepsis (7). Superfluous HMGB1, secreted by macrophages and monocytes, plays a key role in the pathogenesis of sepsis due to bacterial infections (7), as well as in cancers (8). Future breakthroughs in sepsis treatment may depend on addressing the issue of the production of late inflammatory cytokines, such as HMGB1 (9).

As non-coding transcripts of 18–25 nucleotides, microRNAs (miRNAs or miRs) usually target mRNAs to modulate gene expression by 1.2- to 4.0-fold rather than acting as a genetic switch (10). Several miRNAs have been shown to regulate sepsis by targeting different mRNAs. The suppression of the expression of miR-15a/16 in bone marrow-derived macrophages caused by bacterial-derived lipopolysaccharide (LPS) has been shown to increase phagocytosis and bacterial clearance through Toll-like receptor (TLR) 4-associated pathways, subsequently affecting the survival of septic mice (11). miR-147b represents an endothelial barrier of protection against endotoxin-mediated inflammation by decreasing ADAM metallopeptidase domain 15 (ADAM15) expression and attenuating LPS-induced barrier dysfunction during septic challenge by bacterial LPS in endothelial cells (12). Serum levels of miR-122 have been shown to be increased in patietns with sepsis with clotting abnormalities (13). miR-133a levels have also been shown to correlate with the severity of sepsis and, at significantly higher levels, predicted an unfavorable prognosis of critically ill patients (14).

The cholinergic anti-inflammatory pathway (CAP) is a mechanism associated with the vagus nerve. CAP plays a critical role in controlling the inflammatory response by modulating systemic and local inflammation (15). Its molecular mechanisms of action are mediated by the major neurotransmitter, acetylcholine, which activates the nicotinic acetylcholine receptor subunit α7 (α7nAchR). α7nAchR is expressed on macrophages and is an essential component of CAP (3). 3-(2,4-Dimethoxybenzylidene)anabaseine (GTS-21) is an α7nAchR-specific agonist (16). miR-124 reduces early inflammatory cytokine levels in the cholinergic anti-inflammatory response. miR-124 reduces IL-6 production by affecting the signal transducer and activator of transcription 3 (STAT3), and attenuates the release of TNF-α via TNF-α converting enzyme (TACE) (17).

The role of the late inflammatory cytokine HMGB1 as regards its regulation by miRNAs have yet to be reported in sepsis. In this study, using miRNA microarray analysis, we screened 3 target miRNAs: miR-205-5b, miR-196a and miR-193b. Of the 3 miRNAs, miR-205-5b was the most highly expressed followed by decreased HMGB1 levels. Given the important role of HMGB1 in sepsis (18), miR-205-5b is a potential biomarker or therapeutic target in sepsis. In this study, we elucidated the expression of miR-205-5b in different tissues using a mouse model of LPS-induced sepsis and identified the mechanisms of action of miR-205-5b action and its effects on HMGB1.

Materials and methods

Cell culture

The mouse macrophage cell line, RAW264.7, was obtained from the Medical Experimental Center of Zhongnan Hospital of Wuhan University. The RAW264.7 cells were cultured in DMEM with heat-inactivated 10% fetal bovine serum (FBS) (HyClone, Logan, UT, USA) and seeded into 48-well microtiter plates at a density of 2×104 cells/well.

Enzyme-linked immunosorbent assay (ELISA)

A 200 µl supernatant was obtained at 1, 6, 12, 18, 24 and 30 h to confirm the optimal concentration. The levels of TNF-α and HMGB1 in the supernatant were determined by ELISA using a specific kit (eBioscience, Inc., San Diego, CA, USA) according to the manufacturer's instructions with a microplate reader (RT-2100C; Rayto, USA).

miRNA microarrays

Following culture of the RAW264.7 cells for 24 h, cells in the LPS group were stimulated with 50 ng/ml LPS (Sigma-Aldrich, St. Louis, MO, USA), those in the GTS-21 group with 50 ng/ml LPS and 8 µg/ml GTS-21 (Abcam, Cambridge, UK) at 24 h, along with the control group (untreated cells) on an Affymetrix miRNA 3.0 technology platform, which contained 1,111 mature mouse miRNAs. NanoDrop ND-2100 (Thermo Fisher Scientific, Waltham, MA, USA) was used with Agilent 2100 (Agilent Technologies, Inc., Santa Clara, CA, USA) to identify and quantify the total RNAs and ensure their integrity.

Total RNA was tailed with with poly(A) polymerase and labeled with biotin, followed by hybridization, washing, staining and scanning with Affymetrix Scanner 3000 (Affymetrix, Inc., Santa Clara, CA, USA) to obtain the raw image. Raw image processing was performed using Affymetrix GeneChip Command Console software (version 4.0), read by Expression Console (version 1.3.1) (both from Affymetrix, Inc.) and RMA normalization. GeneSpring NGS (version 12.5) (Agilent Technologies) was used for subsequent data analysis. Probes meeting at least one out of two conditions and containing flags in 'P' were selected for further analysis.

Differentially expressed miRNAs were then identified through fold change. The threshold set for upregulated and downregulated genes was a fold change ≥2.0 and p≤0.05. Target genes of differentially expressed miRNAs that represented the intersection of the 3 databases (TargeScan, PITA and microRNA.org) were selected as candidate target genes for further analysis. GO analysis and KEGG analysis were used to determine the roles of these target genes. Unsupervised hierarchical clustering was performed to show the distinguishable miRNA expression pattern among samples.

Experimental animals

We purchased 5-to 6-week-old BALB/c mice (weighing 20–22 g) from the Experimental Animal Center of Wuhan University, an accredited SPF facility according to national and institutional guidelines. The animals were bred under standard conditions with a temperature of 22°C with a 12-h light/dark cycle and had free access to water and a standard rodent diet. In addition, the animals received humane care in compliance with the Principles of Laboratory Animal Care of Wuhan University. All protocols were approved by the Wuhan University of Science and Technology Animal Care and Use Committee (certificate number: TY20120025).

Experimental protocols

The experimental animals were randomly divided into 3 groups as follows: the control, LPS, and GTS-21 treatment (LPS + GTS-21) groups, with 6 mice in each group.

In the LPS group, a single intraperitoneal injection of 15 mg/kg LPS was used to induce peripheral inflammation. The mice in the control group were treated similar to those in the LPS group, but received saline instead of LPS. In the GTS-21 treatment group, 15 mg/kg LPS and 4 mg/kg GTS-21 were injected into the mouse abdominal cavity.

Twenty-four hours after the intraperitoneal injection, blood, lungs, livers, colons and spleens were harvested immediately after the mice were euthanized. Blood was centrifuged at 3,000 rpm for 10 min at 4°C, to obtain the supernatant for further testing. Other tissues were stored in TRIzol for later use.

Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) for the determination of miRNA and mRNA expression

Total RNA extracted from the lung, heart, liver, spleen, gut and renal tissues of the septic mice using TRIzol reagent (Invitrogen, Carslbad, CA, USA), and was reverse-transcribed into cDNA using the RevertAid First Strand cDNA Synthesis kit (Thermo Fisher Scientific). qPCR for TNF-α and β-actin was performed using the 7300 Real-Time PCR system (Applied Biosystems Life Technologies, Foster City, CA, USA). The primer sequences were synthesized for TNF-α (Invitrogen). RT-qPCR was performed in triplicate and data were calculated using the 2−ΔΔCT method, where ΔΔ Ct = ΔCt - ΔCtNC group, ΔCt = CtmiR-205 - CtU6. The primers used for RT-qPCR are listed in Table I.

Table I Primers used for RT-qPCR.

Table I

Primers used for RT-qPCR.

Primers	Sequences
U6-S	5′-CTCGCTTCGGCAGCACA-3′
U6-A	5′-AACGCTTCACGAATTTGCGT-3′
M-miR-205-5p-RT	5′-CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGCAGACTCC-3′
M-miR-205-5p-S	5′-ACACTCCAGCTGGGTCCTTCATTCCACCGG-3′
Universal primer	5′-TGGTGTCGTGGAGTCG-3′
M-HMGB1-S	5′-CGGATGCTTCTGTCAACTTCTC-3′
M-HMGB1-A	5′-GTTTCTTCGCAACATCACCAAT-3′
M-β-actin-S	5′-GTGACGTTGACATCCGTAAAGA-3′
M-β-actin-A	5′-GTAACAGTCCGCCTAGAAGCAC-3′

Transfection of miRNAs

The RAW264.7 cells were seeded into 6-well plates (5×104 cells/well) and cultured in 2 ml Dulbecco's modified Eagle's medium (DMEM) with 5% FBS, maintained at 37°C in the presence of 5% CO2. miR-205 mimics, inhibitor and negative control were synthesized by GenePharma (Shanghai, China). When the cells reached 70–90% confluency, they were transfected with miR-205 mimics or inhibitor, respectively according to the manufacturer's instructions with adherent cell transfection procedures. A negative control of Lipofectamine 2000 (Invitrogen) was maintained as well.

Western blot analysis

In order to obtain the total cellular lysates, the transfected cells and controls were lysed in ice-cold cell lysis buffer (Wuhan Goodbio Technology Co., Ltd., Wuhan, China). The protein concentration was determined using the Bradford protein assay kit (Bio-Rad Laboratories, Inc., Hercules, CA, USA) using a c-globulin standard curve. Proteins were separated by standard SDS-PAGE and transferred onto PVDF membranes (Millipore Corp., Billerica, MA, USA). Non-specific binding sites were blocked using 5% dry skimmed milk, 0.2% Tween-20 in PBS (pH 7.4) and then incubated with HMGB1 antibodies (Boster Inc., Wuhan, China) overnight at 4°C. The PVDF membranes were decolorized 3 times for 5 min each on a shaking bed with TBST at room temperature. The secondary antibody (Wuhan Goodbio Technology Co., Ltd.) was diluted (1:3,000) with TBST, and incubated for 30 min. The PVDF membranes were washed with TBST 3 times for 5 min again. After full access to substrate working solution, developments and fixing, the gray scale was analyzed using the Alpha software processing system (Alpha Innotech Corp., San Leandro, CA, USA).

Statistical analysis

All data are expressed as the means ± standard error of the mean and analyzed using GraphPad Prism version 6.01 (GraphPad Software, Inc., San Diego, CA; USA). Statistical signifiance was determined using a Student's t-test, and a value of p<0.05 was considered to indicate a statistically significant difference.

Results

Optimal LPS and GTS-21 concentration

In order to select the optimal concentrations of LPS and GTS-21, we designed the LPS and GTS concentration gradient. The levels of TNF-α and HMGB1 in the cells were examined by ELISA in order to determine the optimal concentration of LPS needed to induce maximal levels of TNF-α and HMGB1 and the optimal concentration of GTS-21 needed to reduce the levels of TNF-α and HMGB1. Following incubation for 24 h at 37°C, we stimulated the RAW264.7 cells with LPS using the 30, 50 and 80 ng/ml concentration gradients following pre-treatment with GTS-21 using the 6, 8 and 10 µg/ml concentration gradients half an hour earlier. The results suggested that macrophages stimulated with LPS 50 ng/ml expressed more TNF-α and HMGB1 than the other groups; thus, this concentration was considered ideal (Fig. 1A). When the macrophages were stimulated with 50 ng/ml LPS and 3 different concentrations of GTS-21, the decrease in the expression of TNF-α and HMGB1 was greatest at 8 µg/ml of GTS-21 (Fig. 1B). The optimal GTS-21 concentration for treatment was 8 µg/ml, as this concentration achieved optimal inhibitory effects.

Figure 1 GTS-21 inhibits cellular tumor necrosis factor (TNF)-α and high mobility group box (HMGB)1 release. The optimal dose of LPS is 50 ng/ml, and that of GTS-21 is 8 µg/ml. (A) Time courses of TNF-α and HMGB1 induction in RAW264.7 cells stimulated with LPS 30, 50 and 80 ng/ml. (B) Time courses of TNF-α and HMGB1 induction in RAW264.7 cells treated with LPS 50 ng/ml and different concentrations of GTS-21 (6, 8 and 10 µg/ml).

Upregulated miRNA targets in microarray analysis

In the heatmap shown in Fig. 2, each row represents a single miRNA, and each column represents a plasma sample. The legend on the right indicates the miRNA representing the corresponding row. The relative miRNA expression is displayed according to the color scale: red indicates upregulation; green indicates downregulation, as previously described (19).

Figure 2 Heatmap showing 3 differentially expressed (fold change ≥2.0) miRNAs from homogeneous RAW264.7 samples. The numbers with (0+0) indicate the control; numbers with (0+50) indicate LPS-stimulated cells, numbers with (8+50) indicate GTS-21-treated cells.

Differentially expressed miRNAs were identified through fold change (FC). Among the different miRs (miR-205-5b, miR-196a and miR-193b), miR-205 was highly expressed in the GTS-21 group, and its levels were 18.478138-fold those of the LPS group, and 14.27163-fold those of the control group.

Optimal dosage of GTS-21 in mice

The mice (n=6) were intraperitoneally injected with GTS-21 (2, 4 and 6 mg/kg) or saline 30 min prior to LPS administration. The serum TNF-α and HMGB1 levels were analyzed 12 and 24 h later. The ideal dose of GTS-21 was 4 mg/kg as it induced the most prominent inhibitory effects on TNF-α and HMGB1 expression (Fig. 3).

Figure 3 GTS-21 inhibits circulatory tumor necrosis factor (TNF)-α and high mobility group box (HMGB)1 release in endotoxemic mice. The optimal dose of GTS-21 is 4 mg/kg. (A) Serum TNF-α levels were analyzed by ELISA. (B) Serum HMGB1 levels were analyzed by ELISA. n=6 mice per group. Data are the means ± SD; ***p<0.001 vs. LPS-treated group.

Expression profiles of miR-205-5p in mice with sepsis

The expression levels of miR-205-5p targets in the serum, livers, colons, lungs and spleen were quantified by RT-qPCR to verify the upregulation detected in the samples of GTS-21-treated septic mice. The miR-205-5p levels were not similar in the different organs (Fig. 4A–E). The levels of miR-205-5p in the different mouse organs were as follows: serum (LPS vs. control p=0.0033; GTS vs. control p<0.0002; GTS vs. LPS p=0.0001), colon (LPS vs. control p=0.2915; GTS vs. control p<0.0002; GTS vs. LPS p=0.0001); spleen (LPS vs. control p=0.1640; GTS vs. control p<0.0157; GTS vs. LPS p=0.0054); liver (LPS vs. control p=0.8472; GTS vs. control p<0.0015; GTS vs. LPS p=0.0011); and lungs (LPS vs. control p=0.0364; GTS vs. control p<0.0032; GTS vs. LPS p=0.0005) (Fig. 4A–E). In general, treatment with GTS-21 induced a significant increase in miR-205-5p expression.

Figure 4 RT-qPCR analysis of miR-205-5b expression in (A-E) different tissues from BALB/c mice. (F) The serum concentration of high mobility group box (HMGB)1. **p<0.01, ***p<0.001.

The serum concentration of HMGB1 (LPS vs. GTS p=0.0054; LPS vs. control p=0.0022; LPS + GTS vs. control p=0.0063) (Fig. 4F) was significantly increased following the administration of LPS. HMGB1 expression was markedly decreased in serum following treatment with GTS-21 (Fig. 4F).

miR-205-5b attenuates the LPS-induced increase in HMGB1 expression

We first confirmed that transfection with miR-205-5b mimics significantly increased miR-205-5b expression in the cells (Fig. 5A). When the RAW264.7 cells were transfected with the miR-205 mimics, or negative control for 24 h, as shown in Fig. 5, HMGB1 protein expression was altered (p<0.05). Western blot analysis of HMGB1 protein expression revealed that the HMGB1 levels in the group transfected with miR-205-5b mimics were significantly lower than those in the negative control-transfected group (p=0.0280; Fig. 5B and C). However, the miR-205-5b mimics did not affect the HMGB1 mRNA level (p=0.2896; Fig. 5D).

Figure 5 RAW264.7 cells were stimulated with LPS or transfected with the miR-205-5b mimic. (A) miR-205 expression was significantly increased follwoing transfection with miR-205-5b mimic. (B and C) When miR-205 was overexpressed, the protein expression of high mobility group box (HMGB)1 was low in the RAW264.7 cells; (D) however, HMGB1 mRNA expression was not significantly altered in the two groups. *p<0.05, ***p<0.001.

Transfection with the miR-205 inhibitor did not affect the GTS-21-induced suppression of the expression of TNF-α (Fig. 6A); however, the miR-205 inhibitor increased HMGB1 expression 24 h later (Fig. 6B).

Figure 6 RAW264.7 cells were stimulated with LPS or GTS-21 or LPS + GTS-21, or treated with the miR-205-5b inhibitor and stimulated with LPS + GTS-21. The levels of tumor necrosis factor (TNF)-α and high mobility group box (HMGB)1 were determined at 12 and 24 h, respectively. Following transfection with the inhibitor, the RAW264.7 cells were stimulated with LPS and GTS-21, and (A) TNF-α expression was not reduced 12 h later; however, (B) HMGB1 expression was increased 24 h later; n=3 experiments; data are the means ± SD; ***p<0.001.

Discussion

CAP mediates the inhibition of the production of inflammatory cytokines, including TNF-α, nuclear factor (NF)-κB and HMGB1 in sepsis (3). It acts similarly in other inflammatory syndromes, such as arthritis and viral myocarditis (20,21).

The mechanisms of action of CAP entail α7nAchR activation by electrical vagal stimulation or pharmacological agonists. Electrical vagal stimulation is impracticable in sepsis. GTS-21, a selective partial α7nAchR agonist (22), has been proven to effectively reduce the inflammatory reaction and improve prognosis in animals (23–25). Similar to previous studies (22), our findings suggest that GTS-21 suppresses TNF-α and HMGB1 expression in a dose-dependent manner in experiments using mouse macrophages. The curves indicated the optimal concentration of GTS-21 as 8 µg/ml for marophages.

Mammalian miRNAs have emerged as meaningful modulators of cellular function in different organ systems since their discovery (10). Previous studies have shown that miR-124 mediates the cholinergic anti-inflammatory action against early inflammatory factors, including TNF-α maturation by TACE and IL-6 transcription by targeting STAT3 (17); however, it is unknown whether miRNAs affect HMGB1. To elucidate the mechanisms underlying the cholinergic anti-inflammatory activity against one of the most important late inflammatory factors, which correlates with prognosis (7), we performed miRNA microarray analysis to identify potential miRNAs involved. To the best of our knowledge, our study has confirmed for the first time that miR-205-5b expression significantly increased following the activation of α7nAchR in macrophages.

In vitro experiments were more credible as they ruled out the possibility of HMGB1 release from the necrotic tissue. We reported that LPS-stimulated RAW264.7 cells transfected with miR-205-5b mimics displayed a low HMGB1 expression, whereas the expression of TNF-α was not altered. However, transfection with miR-205-5b inhibitor neutralized the effects of GTS-21 on HMGB1 expression. Therefore, our results indicate that miR-205-5b targets HMGB1 expression in the CAP. The results of RT-qPCR demonstrated that miR-205-5b mimics did not affect the HMGB1 mRNA levels, with no significant differences observed between the mimics- and negative control-transfected groups, suggesting a post-transcriptional regulation.

In animal experiments, miR-205-5b expression in the LPS group exhibited a decrease, although this was not significant in the different tissues. This indicates that LPS cannot suppress the expression of miR-205-5b. miR-205-5b targets different proteins in different pathophysiologies (26–29). In this study, we first found that following the activation of CAP, superfluous miR-205-5b suppressed HMGB1 via the transcriptional inhibition of HMGB1 in serum. Decreased HMGB1 expression leads to sharply lowered blood HMGB1 levels.

In conclusion, miR205-5b upregulation suppresses HMGB1 expression in cells and tissues in sepsis. The regulatory mechanisms involved include the suppression of HMGB1 at the post-transcriptional level. miR-205-5b may thus be a potential therapeutic target for the treatment of inflammatory diseases and a possible novel therapeutic strategy against late sepsis.

Acknowledgments

The present study was financially supported by the National Natural Science Foundation of China (no. 30972852).

References